icon-    folder.gif   Conference Reports for NATAP  
  22nd Conference on Retroviruses and
Opportunistic Infections
Seattle Washington Feb 23 - 26, 2015
Back grey_arrow_rt.gif
Bone, Frailty and Vitamin D at CROI 2015
  Todd Brown, MD, PhD
Associate Professor of Medicine and Epidemiology
Johns Hopkins University
This year's CROI had multiple excellent studies presented related to bone health. This review will emphasize those studies with particular clinical relevance or impact on our understanding of the mechanisms regarding bone loss in HIV-infected persons. In contrast to the offerings regarding bone, studies related to frailty/sarcopenia and vitamin D were quite scarce this year. I'll mention these only briefly.
Osteoporosis Epidemiology:
Fracture Risk is Higher in HIV-infected Women:
Anjali Sharma presented a nice analysis regarding the risk of fracture in HIV-infected women compared to HIV-uninfected women using the Women's Interagency HIV Cohort (WIHS) (#775, Report). In this study, 2375 women (1713 HIV+, 662 HIV-) were examined to determine the incidence and risk factors for fracture between 2003 and 2013. A previous analysis in the WIHS had shown a similar fracture rate in the HIV-infected and uninfected women (Yin, AIDS, 2010). With more follow-up time and the aging population, the current analysis showed a clear difference in fracture risk by HIV-serostatus in this young/middle aged population (baseline median age 40 years in HIV-infected and 35 years in the HIV-uninfected). The investigators found that new fractures occurred in 17.5% of the HIV-infected women and 13.6% of the HIV-uninfected women. After multivariable adjustment, the HIV-infected women had a higher risk of fracture compared to the HIV-uninfected women (HR1.30; 95% CI 1.02-1.66). As would be expected with this young population, very few of these fractures occurred at sites typically related to osteoporosis (ie. spine, hip, or forearm). When the analysis was restricted to those who had a fragility fracture (fracture sustained from a trauma equivalent or less than a fall from a standing height), there were no differences by HIV-serostatus.
This analysis was important in a number of ways. First, it specifically examined women. Much of the data regarding osteoporosis in HIV-infected persons, including many of the ART initiation trials (including those discussed below) have had very few women. Second, although the sites where these fractures occurred are not typically related to osteoporosis, it is possible that fractures occurring at peripheral sites like the feet, ankles, and hands may herald fractures at other more central sites (hip and spine), where the clinical consequences are greater. Finally, one of the important predictors of fracture in this cohort was drug use, specifically cocaine use and injection drug use. Whether this is related directly to the drug exposure or other behaviors and comorbidities that accompany illicit drug use (poor nutrition, irregular menses, etc) is unclear. It does tell us that those women with a history of drug use probably deserve more aggressive screening for osteoporosis after menopause.
Changes in BMD Similar by HIV serostatus over 3 years: The UPBEAT cohort from Dublin recruited HIV-infected and HIV-uninfected persons to examine differences in bone mineral density and related factors and clearly found in this younger population that HIV-infected persons had lower spine and hip BMD compared to the uninfected persons (Cotter, AIDS, 2014). In this follow-up analysis (#774, Report), the change in BMD over 3 years was compared in a large subset of the initial cohort. BMD decreased at the total hip and femoral neck in the total cohort, whereas spine BMD remained stable. There was no difference in the change in BMD by HIV serostatus. Interestingly, a factor that was associated with BMD loss in HIV-infected persons was the recent initiation of ART, a finding that is consistent with the known effects of ART initiation on bone. It should be noted however that this population was young (HIV+ 39 yrs, HIV- 43 yrs) and the follow up was relatively short. Longer follow-up in this cohort and examination of older populations is essential to understanding the long-term effects of HIV and its therapy on bone health.
Inflammation Associated with BMD Decline in ART-treated HIV-infected Persons: The SUN is a large multicenter cohort that gathered serial DXAs in a 637 HIV-infected persons over a 4 year period (#772, Report). Over this interval, hip BMD (spine BMD not included) decreased slightly (the magnitude of decrease not described in the poster). The major finding was that higher baseline levels of IL-6 and higher monocyte expression of CCR5 were related to a decrease in hip BMD over the study interval, even after careful multivariable adjustment. This is one of the first demonstrations that inflammation and immune activation may be a driver of bone loss in HIV-infected persons on stable ART. The CCR5 findings are especially intriguing, opening up the possibility that CCR5 inhibition with maraviroc may have a specific impact on BMD (see below). In contrast, memory T-cell phenotypes (CD4+/CD28+ and CD8+/CD28+) were associated with greater increases in BMD. With all of these associations, the magnitude of the effects appears to be small. Nevertheless, the link between inflammation and bone density has been somewhat elusive in longitudinal studies. This study suggests that markers of bone density and bone metabolism should also be probed in studies that examine strategies to decrease systemic inflammation in ART-treated persons.
The FRAX score Underestimates Fracture Risk in HIV-infected Persons: One of the themed poster discussions at CROI addressed cardiovascular risk equations that are used in the general population and the extent to which they underestimate cardiovascular events in HIV-infected persons. It has been expected for some time that there may be an analogous situation for fracture risk. While traditional risk factors for osteoporosis and fragility fracture are the major drivers of fracture risk in HIV-infected populations, factors related to chronic infection with HIV (eg systemic inflammation) and ART also contribute. As a result, fracture risk equations which only take into account traditional risk factors will likely underestimate fracture risk among HIV-infected populations. FRAX (http://www.shef.ac.uk/FRAX/) is one of the major fracture risk equations and is used around the world to assess absolute fracture risk, in order to help make decisions regarding osteoporosis screening and osteoporosis treatment. FRAX calculates a 10 year risk of major osteoporotic fracture (hip, forearm, humerus, and spine) and hip fracture specifically using clinical risk factors. It can be used with or without BMD measurements. Mike Yin and colleagues calculated a modified FRAX score (no history of parental hip fracture was available) in 7064 HIV-infected and 17387 HIV-uninfected men from the Veterans Administration dataset (#141, Report). BMD was not available in this dataset so the FRAX without BMD was used. In both the HIV-infected and HIV-uninfected populations, the modified FRAX score underestimated actual fractures. In the HIV-uninfected group the actual risk of fracture was about 29% higher than the expected risk based on FRAX. In the HIV-infected population, however, the actual risk of fracture was about 62% higher than expected.
What I thought was very interesting and clinically useful about this analysis is that the investigators also showed a way that this inaccuracy can be corrected. In the FRAX, there is a "secondary cause" button that should be checked off if any one of the following risk factors is present: type I diabetes, osteogenesis imperfecta, untreated long-standing hyperthyroidism, hypogonadism or premature menopause (<45 years), chronic malnutrition, or malabsorption and chronic liver disease. This provides a "fudge factor" that accounts for the increased fracture risk associated with these conditions. HIV-infection is not included among these. However, this analysis showed that if this button is checked off, the ratio of observed/expected fractures in the HIV-infected population becomes identical to that which was observed in the HIV-uninfected population, suggesting that this secondary cause button introduces the appropriate "fudge factor" that accounts for increased fracture risk among HIV-infected persons. There is one caveat however about the secondary cause button. It only works when the FRAX without BMD is used (it assumes that all of the effect of all of the conditions on fracture risk is mediated by low BMD). So, this "secondary cause" strategy will not be effective if the FRAX equation is used with BMD. This is how the FRAX is used in the US: to identify a population of patients with osteopenia in whom bisphosphonates are recommended. Clearly, this avenue of investigation requires further exploration. It will be important to know the extent to which FRAX with BMD underestimates fracture risk.
Antiretroviral Effects on Bone: TDF and Alternatives to TDF
Tenofovir disoproxil fumarate (TDF) can be bad to the bone. The biggest bone stories of 2015 CROI were studies that examined the effects of TDF-free regimens on bone mineral density. Most prominent among these were three studies that examined the effect of a tenofovir pro-drug, tenofovir alafenamide fumarate (TAF). Paul Sax presented the combined 48 week results of the two major TAF Phase III ART initiation trials (#143LB, Report). A combined 1,733 ART-naïve, HIV-infected persons were randomized to either TDF/FTC/elvitegravir (EVG)/cobicistat (c) or TAF/FTC/EVG/c. The population was young (median age ~34) and mostly male (85%). Participants randomized to TDF/FTC/EVG/c had the expected drop in BMD over the first 48 weeks (~3%), whereas the decrease in BMD in the TAF-containing arm was about half as much (-1.30 % hip, -0.66% spine). Those in the TAF arm had smaller increases in bone turnover markers and parathyroid hormone compared to those in the TDF arm. As has been previously reported from this CROI and also seen in the phase II TAF trials, the virologic efficacy was excellent in both arms of the study. This study suggests that TAF may be a good alternative to TDF in HIV-infected persons with a higher fracture risk, either because of low BMD, older age, or the presence of other clinical risk factors. I would have liked to see more subgroup analyses to examine whether the benefit of TAF vs TDF is more or less pronounced in certain subgroups (egs, baseline low BMD, older age, younger age (before peak bone mass is achieved), women, race). These analyses may guide clinical decisions about who may most benefit from initiating TAF from a bone perspective.
A single arm, switch study was also presented which examined the bone and renal effects of switching to a TAF-based regimen in HIV-infected persons on suppressive ART with a low glomerular filtration rate (#795, Report). In this study, GFR remained stable, urine protein excretion decreased, and markers of proximal tubule function improved. From a bone perspective, what was noteworthy about the study was that spine BMD increased by 1.9% and hip BMD increased by 0.9% over the 48 weeks. The baseline BMD or T-scores were not reported, so it unclear how similar the study population was to the typical low BMD population in whom a switch to another ART agent would be considered in clinical practice. It would have also been helpful to know whether certain subgroups may show a particular benefit (those with phosphate wasting, low BMD at entry to the study, etc). Nevertheless, this study showed that TAF might be a good alternative to TDF for those in whom bone health is an important consideration.
The third TAF study examined the effect of ART initiation (TDF/FTC/EVG/c) in 23 ART-naïve adolescents in Africa (#953, Report). This study showed an improvement in BMD with ART initiation with a TAF-containing regimen (+2.8%) over 24 weeks. While these findings provide some reassurance since BMD generally decreases with ART initiation, without a control group (either a group initiating another ART regimen (perhaps a TDF-containing regimen) or a matched HIV-uninfected group), it is somewhat difficult to put the findings into context. The small sample size is also a limitation. Investigation into the bone effects of TAF in this young population is important since TDF may lead to an attenuation of peak bone mass (at age 25-30) and BMD at age 30 is a major predictor of BMD at age 65-70, an age at which fractures increase exponentially.
Maraviroc as an Alternative to TDF: Other ART medications also may be good alternatives to TDF. In ACTG 5303 (#769LB, poster), we randomized 262 HIV-infected, ART-naïve individuals to either TDF/FTC/darunavir (DRV)/ritonavir (r) or MCV/FTC/DRV/r and examined the change in BMD over 48 weeks. The BMD decline was less in the MVC arm vs the TDF arm at both the hip (-1.51% vs -2.40%) and the lumbar spine (-0.88% v -2.35%). It should be noted that the dose of MVC was 150 mg daily and therefore should be considered experimental.
Both regimens were well tolerated and had similar virologic efficacy. It is unclear if the same results would have been seen if MVC had been given at the recommended dose. Nevertheless, the findings may suggest that MVC may be a good alternative to TDF in those with increased risk of fracture initiating ART.
Other Alternatives to TDF- A Minimalist Approach: In the MODAt trial (#773, Poster), 69 HIV-infected individuals with virologic suppression receiving ATV/r 300/100mg+2 N(t)RTIs were randomized to either ATV/r monotherapy or to continue on their current regimen. The majority of participants (87%) were receiving TDF at baseline. At week 48, those who were taking ATV/r monotherapy had a greater percentage increase in lumbar spine compared to those who remained on their NRTIs +ATV/r (+1.8% vs -0.99%;p=0.002), with similar trends at 96 weeks (+0.83% vs -0.81%; p=0.075). At the hip, similar results were seen favoring ATV/r at 48 weeks (+1.3% vs -0.20%;p=0.069) and at 96 weeks (+1.3% vs -0.47%; p=0.071), although these results were not statistically significant. These findings suggest that from a strictly bone perspective ATV/r monotherapy may be a good strategy for those at risk of fracture with TDF, but the safety of this approach from a virologic perspective has not been clearly demonstrated. Indeed, in this MODAt substudy, approximately 25% (10 of 39) who were randomized to ATV/r monotherapy and had a 96 week DXA were excluded from the bone analysis because the required re-intensification of their ART regimen. Even I know (i.e someone who does not treat HIV) that this result is not good from an HIV control standpoint.
Possible Mechanisms of TDF effect on Bone: The mechanisms underlying TDF's effect on bone are not clear. One hypothesis is that TDF induces secretion of PTH which in turn increases bone turnover. Indeed, in ART initiation trials (including the TAF trial above), PTH increases with TDF and the increase is greater than that seen with comparator drugs. Havens et al suggested that this elevation in PTH may be due to a functional vitamin D deficiency, whereby TDF-induced increases in vitamin D binding protein lead to reduced free 1,25 dihydroxyvitamin D concentrations (Antimicrobial Agents and Chemotherapy, 2013). The lower 1,25 dihydroxy vitamin D then drives PTH secretion. In an interesting CROI poster, Menzaghi et al suggest a more direct mechanism (#765, poster). The secretion of PTH is regulated by the calcium-sensing receptor (CaSR) in the parathyroid tissue. When the receptor is activated (i.e. by small increases in the circulating levels of calcium), PTH secretion decreases. In the kidney, CaSR activation in the proximal tubule leads to phosphate reabsorption. In this study, human embryonic kidney cells were transfected with CaSR wild-type gene and activity of the receptor was assessed in the presence of calcium with or without TDF. The authors found that CaSR activity was reduced in a dose-dependent manner in the presence of TDF. If this were also seen in vivo in the parathyroid gland, the consequence would be increased PTH secretion. Interestingly, this is a similar mechanism by which lithium carbonate induces hyperparathyroidism in treated patients. An argument against the importance of this mechanism in TDF-treated patients is that unlike with lithium, TDF treatment is not associated with hypercalcemia. It is possible however, that the inhibitory effect of TDF on CaSR activity is mostly seen in the proximal tubule, thereby enhancing phosphate wasting. This observation definitely deserves additional study.
In addition, there might be genetic factors that might predispose certain patients to the ill-effects of TDF. In a cohort study of 294 TDF-treated, HIV-infected patients in Italy, polymorphisms in the vitamin D receptor were associated with higher PTH levels independent of circulating 25 OH vitamin D concentrations (# 791, poster). Whether this polymorphism also is associated with lower bone mineral density should be investigated further.
Statins and Bone in HIV: A Nail in the Coffin: Last year, Grace McComsey presented the 48 week results of the Stopping Atherosclerosis and Treating Unhealthy bone with RosuvastatiN in HIV (SATURN-HIV) trial, showing that rosuvavstatin treatment was associated with increases in hip bone mineral density in ART-treated HIV-infected persons with evidence of heightened inflammation. This year, the 96 week results were presented and found that the apparent benefit at 48 weeks of rosuvastatin was not sustained at 96 weeks (#771, poster). These data, along with the null results of the JUPITER study on fracture which were also published this year (Pena, JAMA Int Med, 2015) and a host of other null studies, suggest that statins (at least the ones that are currently available) likely do not have a role in reducing fracture risk, as was once previously hypothesized.
Quantitative Ultrasound: A Good Imaging Modality in Resource Limited Settings?: DXA machines are expensive and are not readily available in many parts of the world. Quantitative ultrasound (QUS) may offer a cheaper alternative to evaluate fracture risk, has been shown to predict fractures in multiple cohorts, and may provide information that is complementary to DXA (see Guglielmi, Radiology Clinics of North America, 2010 for an excellent review). In a study of 244 HIV-infected patients, Pinzone (#776, poster) showed that the quanitative ultrasound index (QUI) differed from HIV-uninfected controls and was correlated with known fracture risk factors in HIV-infected persons (age, TDF use, PI use). The authors suggest that use of QUS before DXA screening could eliminate the need for DXAs in 44% of those who would be eligible for DXAs based on clinical risk factors. The caveat for this study is the lack of concurrent DXA data to get a better understanding of the correlation between DXA and QUS in an HIV-infected population and as well as the accuracy of this "QUS first" approach.
Two other studies also used QUS. In a study of ART-naïve, HIV-infected women in Rwanda (#780, poster), QUS at the radius was similar to an HIV-uninfected control group. In a study on children in South Africa (#932, abstract), HIV+ children (n=206) were found to have lower QUS indices compared to HIV-uninfected controls. These studies suggest that QUS is feasible for research studies, and potentially clinical practice in resource limited settings.
Frailty and Vitamin D-Meager Offerings:
In contrast to the studies on bone, the studies investigating frailty and vitamin D were not plentiful at this year's CROI. Damani Piggott (#783, poster) found in the ALIVE cohort that the frailty phenotype was associated with AIDS and non-AIDS hospitalization, independent of CD4 cell count or viral load, providing further evidence of the clinical utility and predictive value of this phenotype. In the MACS (#786, report/poster), we found that self-reported balance confidence was a stronger predictor of falls compared to measures of physical function, suggesting that it is important to ask patients about how confident they are about staying on their feet.
The vitamin D studies at this year's CROI were especially scarce. There has been a huge explosion of vitamin D studies in the last 5 years both in HIV-infected and HIV-uninfected populations. Most of these studies have been observational studies relating vitamin D deficiency to various outcomes. What is now needed from a clinical perspective is to determine whether replacement of vitamin D will reduce clinical outcomes. Fortunately, these studies are underway.